INDEX
Explanations
song titles, particularly those with a focus on emotions or storytelling
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.15
0.5%
394
+0.13
0.4%
2019
+0.12
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1826
+0.15
0.01
50
+0.13
0.02
752
+0.12
0.02
Negative Logits
FORMANCE
-0.65
flanges
-0.63
bituminous
-0.58
nozzles
-0.56
bushing
-0.56
aislada
-0.56
CLUDING
-0.55
resistive
-0.55
ductile
-0.54
bulkhead
-0.54
POSITIVE LOGITS
confé
1.28
vété
1.14
marchand
1.02
Secrétaire
0.97
trouva
0.96
monstre
0.94
puits
0.94
héro
0.94
maît
0.94
rempliss
0.93
Activations Density 0.097%